Thirty-one patients with osteoarthritis of the knee were treated with either diflunisal (n = 17) or naproxen (n = 14) in a 12-week open-label study. Treatment was begun with 500 mg BID of diflunisal or 375 mg BID of naproxen. Patients not showing an adequate response to these dosages were given increases to 750 mg BID of diflunisal (n = 7) or 500 mg BID of naproxen (n = 8). Both drugs achieved statistically significant improvements in pain indices, tenderness, swelling, morning stiffness, functional capacity, knee flexion, and 50-foot walking time, and no significant difference was found between the two drugs. At the end of the study, all patients taking diflunisal and 11/14 patients taking naproxen felt that they had improved with treatment. Drug safety and tolerability were assessed in 21 patients given diflunisal and 16 given naproxen (including patients not part of the efficacy evaluation). Six (29%) patients in the diflunisal group and four (25%) in the naproxen group experienced side effects; three were withdrawn from the diflunisal group and one from the naproxen group because of adverse effects. In general, both drugs were well tolerated.